Build one level of the tree by clustering and summarizing nodes.

Evaluate a mathematical expression Args: expression: Mathematical expression to evaluate Re

Vectorized calculation of expected scores for multiple matches. Args: ratings_a: Array of ratings for model A ratings_b:

Calculate BM25 score for a single term in a document

Call an MCP tool by spawning a fresh server connection Args: tool_key: Tool key in format "server.tool_name"

检查 API Key 配置

Check if all dependencies are available

Check if all dependencies are available

Check if the environment is properly set up

Check if environment is properly configured

Check if required environment variables are set

Check if environment is properly configured

Check if required environment variables are set

Check if local retrieval pipeline is running

Check system compatibility

Choose an action using LLM reasoning with in-context learning.

Choose action (simplified for demonstration).

Smart chunking that respects paragraph boundaries with soft and hard limits. Strategy: - Accumulate paragraphs until

Split text into chunks with overlap.

Split text into chunks with overlap.

* Clean up old temporary files

Clean up existing indexes and document store

Clean up existing indexes and document store

Clean up test knowledge base.

Clear the page cache

Clear the index before demo

Close browser session.

Cluster embeddings using Gaussian Mixture Model.

Execute Python code for complex calculations and data processing Args: code: Python code to execute

* Collect streaming response data

比较搜索 - 搜索并比较多个项目 Args: items: 要比较的项目列表 aspect: 比较的方面 Returns: 比较结果

Compare results from different backends for the same query

Compare different extraction modes for the same content

Compare different KV cache implementations Args: api_key: API key for Kimi task: Task to execute root_dir: Root

Create visualization comparing learning curves of both methods.

Compare different models for the same task

Compare agentic vs non-agentic performance

Compare agentic vs non-agentic performance

Compare three models on the same prompt

Compare contextual vs non-contextual retrieval

Compare streaming vs regular responses

Compress content for message history (used in windowed context strategy) Args: content: Content to compress

Compress search results based on the selected strategy Args: search_results: Raw search results from web tool

Compute aggregate metrics from results

Compute aggregate metrics from results

Compute leaderboard using Bradley-Terry model (official Chatbot Arena method). Args: df: DataFrame with columns 'model_a', 'mode

Compute Elo ratings using Bradley-Terry model with Maximum Likelihood Estimation. This is the official method used by Chatbot Arena for

Compute leaderboard using online Elo updates (K=4, official value). This method updates ratings sequentially as matches are processed.

Configure RAG and evaluation settings

Configure system settings

Connect to an MCP server with proper error isolation NOTE: This stores tool metadata only. For actual tool execution,

Add contextual information to conversation chunks. Args: chunks: List of basic conversation chunks f

Convert currency using live exchange rates Args: amount: Amount to convert from_currency: Source cur

Estimate token count (rough approximation)

* Create a test audio buffer with spoken content * This creates a simple sine wave that simulates audio content

* Create a sample audio buffer for testing * This creates a simple sine wave audio buffer

Create a flow diagram showing attention evolution over generation steps Args: attention_steps: List of attention step dictionari

Create a heatmap visualization of attention weights Args: attention_weights: 2D list of attention weights [output_len x total_le

Create a demo conversation with important information

Create a demo conversation with important information

Create a screenshot of Excel file (requires GUI environment). Note: This is a simplified implementation that exports to image. Args:

Create a sample financial report PDF for testing

Create initial messages with proper format for Qwen3

Create a task requiring multiple currency conversions and calculations Returns: Task description

Create sample files for testing

Create a sample Intel architecture documentation for testing.

Create sample queries for testing.

Create a simpler expense report for quick testing

Create a summary visualization showing tokens and their attention patterns Args: result: Generation result dictionary sa

Create data for visualization (can be used with plotting libraries)

Delete a memory card. Args: category: Card category card_key: Card key Returns:

Delete a document from the store.

Delete a document.

Delete a document from the store. Args: doc_id: Document ID Returns: True if delete

Delete lines from file. Args: file_path: File path start_line: Start line number (1-indexed, inclusi

Interactive demo with proper tool calling

Demonstrate basic Elo rating calculation with synthetic data.

Demonstrate basic system hint features

Demonstrate the improved category evaluation selection

Demonstrate the improved category evaluation selection

Demonstrate code generation and analysis capability

Demonstrate using both tools together

Compare with and without system hints

Run a side-by-side comparison

Demonstrate a conversation with multimodal tools

Demonstrate the separated memory system architecture

Run a quick demo with sample PII-containing text

Demo streaming with Ollama backend

Demonstrate the context loss problem